Jump to content

How would you use sensors to detect snow?

Recommended Posts

Hi everyone, i've been searching all week for a solution but nothing has come yet. Being a beginner didnt help either :)


Im doing a project and would like to keep it simple while providing upgradability.

I made a heater which is used to melt snow. Power is about 800-1000 watt on a 120v outlet.


The idea would be to detect when its snowing and start slowly heating to prevent snow buildup. When there are no more snow the heater would stop and the controller return to standby. Maybe probe each hour for snow.


I couldnt find what would be the best way to do this. The sensor would need to be able to operate in temperature as low as -30 F.

A temperature sensor is off course needed and I was thinking that maybe 2 ultrasonic sensor would detect snow accumulation while not having to be exposed to direct weather.


I am purchasing components soon with a new launchpad so i would like to save on shipping. It would be amazing to have some advice on what I might need. Im not sure either about the relay.


Thank you for taking the time to read this,

Link to post
Share on other sites

Surprisingly i didnt find that link with all those hours spent on google. It looks like a great idea, but for the data to be valid the sensor need to at a steady place and will be under quite a lot of stress in storms. Its gonna be in front of the house btw.

Thank you i will learn a lot with your link.

Link to post
Share on other sites

You know, if you come up with a good solution, you might want to look into patenting it.


LED traffic control lights are all the rage with cheapa.. er frugal cities, but they suck during the winter. Most solutions involve just sticking a heater on them full time in the winter. It's not an intelligent system.

Link to post
Share on other sites

Do you need to detect accumulating snow or just the fact that it is snowing?


There are 2 ways I can think of. They aren't super complicated, but they would require a bit of work to tweak into a working instrument.


1. Hot wire probe (or king probe). This type of instrument has a thin wire heated above 100C. When water or snow hits the wire, it cools off the wire which changes its resistance. The wire is part of a balanced bridge that increases voltage to the wire in response to the change in resistance. You can measure the voltage supplied to the wire and determine if rain/snow is hitting it. If you build it well, you can even back snow/rain rate based on the voltage. These things are still used on aircraft to measure total liquid water in clouds. Here's a link to a description and some plans: http://journals.ametsoc.org/doi/pdf/10. ... 2.0.CO%3B2


2: Measure the albedo of a test target. Take two small plates, paint them black and place them out in the open. Take a pair of photo diodes and shield them to limit their angular detection range and arrange them to "look at" the plates. Add some illumination diodes to work at night. Now heat one of the plates so that it stays above 0C. The heated plate will serve as the reference. By reading the difference in apparent brightness between the two plates, you can determine if snow is building up on the unheated plate. It would be a good idea to have a circuit that would heat the other plate once it hits some reflectivity threshold, then you could tell when snow starts and stops, and gauge snow fall rate based on the duty cycle of the heater.

Link to post
Share on other sites

maybe. would depend on the expected environmental conditions and set up, i guess, some of which may be controlled, some not.


i think i imagined a relatively large surface area, at least a foot by a foot, say, where from 'mass readings' could be taken periodically and averaged over an hour or so and compared to a current mass and temperature reading maybe. would be less power consumption than applying constant heat. once it was decided that heat may need to be generated (to melt the snow) other methods as suggested above could be employed too. t'was just an idea.


*edit* in fact i imagined an umbrella at first - why isn't a simple and robust umbrella style construction (no micro required) used to protect traffic lights from the elements? it would sit above the lights which are generally viewed from below. or have i, once again, missed an important point or three? anyways, an umbrella like structure would allow rain to flow off and snow to collect, adding weight which could be measured at the central pivot.

Link to post
Share on other sites

Thank you for all the insights. The goal is to detect when there is snow, no need to measure how much there is.

I think that the hot wire probe seems like a good idea. Looks like there are a lot of ways to do it. What about a a color reading sensor looking at a black surface? it would for sure detect when there is white snow covering the black area? I dont know however if such sensor exist in weatherproof version, or if it would work behind a small glass.

Link to post
Share on other sites

I think you need to give practical boundaries to your problem so that it becomes solvable.


What is being heated? Indoors or outdoors?

What size of area is being monitored for snow?

Is it a driveway, a parking garage ramp, a building roof?


If you can limit the size then you can limit the amount of sensor data and craft a heater control algorithm and know that it's going to succeed.


If you said "I need to keep the stairs clear of snow" then you have a natural size limitation. Then you could start to analyze the variables that change when snow is falling onto that area.


If you have to monitor a specific set of stairs (say at school or at home) then you could analyze the surrounding structures to see what doesn't change. Then you could figure out where you could place sensors. For example, you could place some sort of beam breaking sensor to sense that something is falling in front of it.


If you have a finite area to monitor and to heat then you could sense the added mass of snow, the presence of moisture, the temperature, the wind speed and the humidity in the air. Based on those inputs you could calculate the dew point and assess whether or not that temperature is below 32'F (0'C).


Armed with the dew point, the temperature, the humidity, the dynamic mass (extra mass in the monitored zone) you could then make an educated guess of when to turn on/off the heater.


If the monitored area is indoors then the problem becomes a lot easier. If the temperature is below freezing then turn on the heater. Keep it on until the temperature rises to a specific set point then turn it off. Allow the unit to cycle off and on. Monitor the temperature and make sure that the sinusoidal temperature readings are above the 32'F (0'C) threshold at all times.


If the area is a building roof and your device is going to keep the snow from building up then just turn on the heater below 32'F and turn it off above 34'F.



Can you describe your snow monitoring challenge a bit more?

Link to post
Share on other sites

The project goal is to build low cost snow melting carpets from recycled rubber. 7 stairs which are 8 feets wide but the carpet only need to cover 3 feet. The goal is to provide a safe and practicable stairway during the winter for a friend who cannot use a shovel because of a disability. Of course this project will have a lot of other useful applications.



Link to post
Share on other sites

if that's the case, then detecting snow isn't good enough.... What will happen is the heater will stop when the snow stops falling, and the mat will turn into a big sheet of well groomed ice. I think you are going to need to keep the mat slightly above 0C constantly (or during certain defined operating hours).

Link to post
Share on other sites

In that case, weight sensors. More than a x% increase in weight over n minutes (reasonable amount of time to rule out people walking on the stairs). Place one or two on each step, and only trigger when a few have passed the threshold weight. Light sensors as a secondary sensor.


actually, read the different product descriptions on this site:


Some commercial products, always good for ideas. One uses a heated moisture sensor, basically a high wattage resistor sitting on top of a set of traces. When the snow hits it and melts, the resistance drops from infinity to near zero (exactly how a tv remote control button works), providing a measurable voltage.


Another uses weather sensors. Precipitation + temperature fitting snow weather triggers the heaters.

Link to post
Share on other sites

i would also consider the texture of your rubber carpet tiles - you'd want some sort of positive physical feedback under foot through a thick winter shoe sole before committing full body weight to a step if you were unsure as to whether the way was clear, particularly if vision were a problem... perhaps the heating element could be a sturdy raised metal grid of some sort? thinking cattle grid, not sure why... just thinking out loud again.


*edit* hmmm... if weight were decided upon as the means of detection maybe the floating grid heating element should be employed as detector as rain could collect and apply mass to the rubber mat below but not the floating grid? could this be combined with the hot wire suggestion?

Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Create New...